Earlier Intervention Is Better for Toddlers with ASD: Evidence from a Randomized Controlled Trial of the Early Social Interaction Project

Background: Early diagnosis of autism spectrum disorder (ASD) is on the rise, spurred by advances in early detection methods and recommendations for screening at 18 and 24 months (Johnson & Myers, 2007). Naturalistic developmental behavioral interventions (NDBIs) have been shown to improve language, cognitive, and/or social outcomes for toddlers diagnosed with ASD (Dawson et al., 2010; Kasari et al., 2010; Wetherby et al., 2014). Although there is agreement that intervention before age three is important, empirical evidence for the notion that “earlier is better” is limited and comes almost entirely from observational studies (Granpeesheh et al., 2009; Harris & Handleman, 2000). There is a critical need for experimental treatment studies that test timing effects to guide the optimal age of intervention for ASD. 

Objectives: To report on findings from a randomized controlled trial (RCT) designed to test timing effects of early intervention for toddlers with ASD under age three. This study compared improvement during Early Social Interaction (ESI) between toddlers randomized to receive individualized treatment early (~18 months) or 9 months later (~27 months). 

Methods: A complete crossover RCT design was employed in which toddlers (N=82) received ESI-Individual at either 18 or 27 months of age. At 18 months, toddlers were randomly assigned to either ESI-Individual or ESI-Group as their first treatment condition. After 9 months, children received the other treatment condition for an additional 9 months, totaling 18 months of treatment. Child outcome measures were administered at baseline, crossover, and end of treatment: Communication and Symbolic Behavior Scales (CSBS), Mullen Scales of Early Learning (MSEL), Vineland Adaptive Behavior Scales, Second Edition (VABS-II), and Autism Diagnostic Observation Schedule (ADOS). 

Results: Generalized linear models revealed that toddlers who received ESI-Individual at 18 months showed greater gains during this treatment than those who received ESI-Individual at 27 months. These timing effects were demonstrated for receptive language (MSEL and VABS-II), expressive language (VABS-II), social communication (CSBS), symbolic skills (CSBS), and daily living skills (VABS-II). Timing effects were not observed for autism symptoms (ADOS), motor skills (MSEL and VABS-II), or visual reception skills (MSEL). 

Conclusions: These findings extend current knowledge on efficacy of NDBIs by demonstrating that ESI-Individual treatment leads to greater gains when initiated at 18 months compared to 27 months. This study represents the first effort to use RCT methods to rigorously test whether earlier is better for toddlers with ASD, allowing for strong causal conclusions to be made. This study is also the first to examine timing effects under age two. Previous studies have generally examined the effect of treatment started by age three, but this study examined even younger children in order to determine the optimal age for treatment. Results provide strong support for treatment beginning at 18 months and suggest that even a narrow window of 18 versus 27 months may have a critical impact on early intervention. These findings also underscore the importance of screening for ASD early in the second year of life and referring toddlers to autism-specific intervention programs by 18 months.